RIETI-NISTEP Policy Symposium
Open Innovation as a Key Driver of
Japan’s Industrial Competitiveness
Handout
Ashish ARORA
Professor, Fuqua School of Business, Duke University
Research Associate, NBER
August 21, 2015
Research Institute of Economy, Trade and Industry (RIETI)
http://www.rieti.go.jp/en/index.html
The Acquisition and
Commercialization of Invention in
the American Economy
A. Arora, W. Cohen (Duke and NBER),
J. Walsh (Georgia Tech)
Funded by the
National Science Foundation
Ewing Marion Kauffman Foundation
The paper
 Broad-based evidence on
the rate of innovation in the
U.S. and the extent to which
it relies upon external
invention
 5,000 + American
manufacturing firms.
 Innovation v Invention
 Sources and channels
 Sample frame: Dun and
Bradstreet, stratified by industry
(28 3-4 digit NAICS), firm size (6
strata)
 Phone survey: marketing
Motivation
 External sources of
invention
 Neglected till recently
 Gains from trade and
specialization
 Existing literature for
U.S. has studied of
specific channels (e.g.,
licensing, coop), or
specific sources (e.g.,
customers, universities).
 No CIS available for U.S.
managers or business manager
2
 Rate of innovation and
Motivation
dependence on external
source co-determined.
 2 Margins: (1) External v
No Innovation and (2)
Make or Buy
Supply and Demand for
invention
 Firms may be at different
demand for
invention
margins  technical
capability of firm
$
 Demand curves = derived
demand for innovation
External
supply of
invention
w
c′(Q) =
internal
supply
 Depends upon product market
competitive conditions …
 External supply curve reflect
ability and number of sources
 Internal supply reflects
efficiency of internal R&D
Qinternal
Q1
Qtotal
Inventions 
3
Project and data
 A survey of product innovation for US manuf. firms in 2010
 + selected services, not included for today
 Population: All firms in these industries, not just innovators
 Focus on “most important innovation” for the “main line of
business”
 Firm = Business, (not parent corporation)
 Sample frame: Dun and Bradstreet, stratified by industry (28 3-4
digit NAICS), firm size (6 strata)
 Oversampled large and startup firms and innovative industries
 Phone survey: marketing managers or business manager (recall
that this is an innovation—not an R&D—survey)
 6685 /22,000 responses (30.3% response rate)
 Today exclude service firms and tiny firms (< 10 employees),
leaving 5,157 in sample.
4
Non-respondent bias tests
 Compared D&B data for respondents and non-
respondents
 Sample is representative of population on:
 Firm age, being multiproduct, region, or likelihood to
export.
 Lower response rates for:
 Large firms, especially Fortune 500 firms (about 20%
response rate)
 Pharmaceuticals also had a low response rate (still over
20%)
 Used Census data to construct industry and size
class post-sampling weights to correct for
response bias
5
Validating Innovation Measures:
Industry Correlations across Measures
External Indicators
ACS
NTF
ACS NTM
BRDIS NTF
.72
.76
Europe-wide CIS NTM
.71
.72
BRDIS R&D Performers
.72
.72
CIS Innovative Activity
.70
.68
BRDIS RDI*
.59
.52
R2 any patent application
(PATSTAT)
.72
.74
R2 patent count
(PATSTAT)
.54
.47
R2 forward citation count
(PATSTAT)
.56
.49
*BRDIS NTF and BRDIS RDI R2 =.35
NTF: In 2009, have you
earned revenue from any new
or significantly improved
goods or services in
[THREAD INDUSTRY]
introduced since 2007, where
“New” means new to your
firm.
NTM: Of all the new or
significantly improved
products or services you
brought to market in
[THREAD INDUSTRY] during
the three years, 2007-2009,
think of the one that accounts
for the most revenue. Did you
introduce this innovation in
your industry before any
other company?
6
Examples of innovations in sample industries
Industry
Food
Food
Beverage
Textile
Textile
Paper
Paper
Petroleum
Chemicals
Pharmaceutical
Pharmaceutical
Plastics
Minerals
Minerals
Metals
Metals
Electronics
Electronics
Semiconductors
Semiconductors
Transport Equipment
Innovation
Antioxidant chocolates
Live active cheddar cheese with probiotics
Vitamin-enhanced flavoured spring water
Heat resistant yarn
New varieties of garments
Low surface-energy tapes resistant to air, water, detergents, UV light
Hanging folder with easy slide tab
Non-detergent motor oil
BioSolvents – water based emulsion technology
Oral gallium to prevent bone decay
Inhalation anaesthetics
Styrene based floor underlayment
Multi-wall polycarbonate recyclable panels
Solar glass and coating technologies solar modules
Solder system & nanofoils
New water faucets and bath products
USB-to-GPIB Interface Adapter
20-h IPS Alpha LCD Panel
Linear voltage regulators
Phase change memory
Improved alcohol sensing system
7
Rates of innovation and imitation, manufacturing industries (wtd)
%
NOSI
% NTM
40
13
38
15
33
8
50
24
63
28
48
16
31
9
38
9
39
10
46
20
76
33
61
27
60
37
54
26
53
27
56
22
48
19
Imitation
% (NOSINTM)
% sales
% sales
from focal
% NTM
from NOSI innovation patented
16
9
24
19
15
51
15
7
11
17
9
42
23
13
61
14
6
42
21
14
35
14
5
23
28
8
35
24
14
52
38
9
58
29
18
59
17
7
54
25
13
53
25
11
34
37
31
72
30
10
45
INDUSTRY
Food & Bev
Text
Wood
Chem
Pharma
Plastics
Minerals
Metals
Fab Metals
Machinery
Electronic
Semi Con
Instrument
Elec Equip
Auto
Med Equip
Miscl
N
362
210
385
365
128
340
323
324
424
384
146
302
135
344
339
136
510
All mfg
5157
43
16
28
22
11
42
Large (> 1000)
1268
66
38
28
26
12
63
945
54
23
32
24
16
47
2944
40
13
27
21
12
Med. Firms
Small (< 100)
27
22
25
26
35
31
22
29
29
26
43
34
23
28
25
34
29
8
36
Overall
 43% report new-to-the-firm (NTF) product introduction
 36% of new products are innovations -- NTM (i.e., about 16% of all firms
innovate and 28% imitate)
 Innovation varies across industries
 Imitation more stable
 42% patented focal innovation
 First estimate of patent propensity for U.S.
 4% firms licensed technology to others without introducing new
products.
 New products account for 22% of revenue for NTF firms; Focal
innovation accounts for 11% of revenue for NTM
 40% of innovators invest in new marketing or sales channels, while 47%
invest in new skills or equipment
 Only 38% of innovations commercialized using existing
complementary capabilities alone
 Lack of complementary capabilities apparently not inhibiting
innovation?
9
SOURCES of
INNOVATIONS
10
External sources
 (for the focal innovation) Did any of the following originate
this innovation, that is, create the overall design,
develop the prototype or conceptualize the technology?
[Responses not mutually exclusive]
Supplier
14%
Customer
27%
Consultant/ Comm. Lab/ Service provider
8%
Independent Inventor
7%
University/Govt Lab
5%
Other Firm in industry
8%
Any external source
49%
11
Sources of invention by industry, percentage shares
INDUSTRY
Food & Bev
Text
Wood
Chem
Pharma
Plastics
Minerals
Metals
Fab Metals
Machinery
Electronic
Semi Con
Instrument
Elec Equip
Auto
Med Equip
Miscl
All mfg
N
Any
Supp Cust
Other Consult./ Ind.
Univ Specialist
External
Firm Serv prov Inventor
(a)
(b)
(c)
(d)
(e)
(f)
(d+e+f)
46
34
15
8
1
5
0
6
63
33
52
102
30
74
36
44
60
98
50
91
53
98
101
36
106
50
52
49
50
53
49
49
48
49
45
62
48
44
52
49
46
32
22
17
2
11
6
29
10
7
11
16
5
12
11
18
8
26
27
15
9
28
23
30
38
36
17
49
26
26
28
22
20
4
11
5
17
5
3
11
6
10
10
9
7
4
12
4
13
3
14
10
6
11
8
11
0
12
8
13
11
8
6
13
10
6
1
3
6
16
12
4
4
7
6
8
9
7
17
9
9
0
0
5
19
4
10
7
3
6
5
9
1
4
15
15
2
9
15
16
30
27
27
13
7
21
14
23
19
17
25
32
18
1127
49%
14%
27%
8%
8%
7%
5%
17%
Did any of the following originate this innovation, that is, create the
overall design, develop the prototype or conceptualize the technology?
[Responses not mutually exclusive]
12
Sources of invention, by firm size
Any
External
Supp
Cust
Other
Firm
Consult./
Serv prov
Ind.
Inventor
Univ
Specialist
Large
(>1000)
51%
22
25
9
7
4
6
14
Medium
(100-1000)
48%
12
26
9
8
4
5
16
Small
(<100)
49%
13
28
8
8
9
5
18
All
49%
14
27
8
8
7
5
17
 No difference in external dependence by size!
 Large firms favor suppliers and universities relative to smaller firms
 Small firms favor independent inventors (9%) relative to large firms (4%)
 2.5% of firms in sample are “startups.” But startups are reported to be
the source of innovation for 13% of firms acquiring inventions externally
13
Innovation and the Sources of Invention
58%
US
Manufacturing
Firms
27%
42%
Innovator Patenting
Propensity
No new
products
6%
Imitation
(new-to-firm)
Innovation
(new-to-mkt)
16%
42%
Source:
51%
Internal
invention 44%
External
invention
49%
40%
External inventor patenting: ~25%
14
External sources and
rate of innovation
Supply and Demand for
invention
demand for
invention
$
 Only 34% of externally reliant
innovators report they could
have obtained from alternate
source  (66% x 49% = 33% of
innovations not happen without
external source)
 Innovation rate drop from 16%
to 11% without external source
 Multinomial assumptions 
External
supply
of
inventio
nw
c′(Q) =
internal
supply
Qinternal
Q1
Qtotal
Inventions 
Innovation drop from 16% to
9% without external source
 Bottom line: Two margins
 Not just “make or buy”
 Weak firms / small firms
more dependent on
external source for
innovation
15
CHANNELS
16
Channels for acquiring innovation overall and by industry (wtd.)
INDUSTRY
Food & Bev
Text
Wood
Chem
Pharma
Plastics
Minerals
Metals
Fab Metals
Machinery
Electronic
Semi Con
Instrument
Elec Equip
Auto
Med Equip
Miscl
All mfg
JV / Coop
Service
N
R&D
M&A
License Contract Informal Market*
26
73
11
17
20
15
32
Market
only*
17
10
23
38
13
32
15
14
22
35
19
37
19
27
37
13
43
76
54
67
40
61
69
64
58
53
83
64
41
62
67
45
63
7
12
6
39
25
13
19
0
11
5
15
6
19
11
15
3
20
5
5
56
9
12
3
9
6
3
15
42
21
33
17
13
17
45
35
8
27
11
17
5
17
19
30
14
36
18
30
20
9
32
29
17
36
56
46
72
42
12
43
14
42
21
31
33
34
50
44
82
45
36
39
14
34
22
37
60
55
53
57
33
16
15
18
43
13
17
18
0
22
3
16
45
22
29
24
14
423
61
10
13
21
37
37
16
17
Channels for acquiring innovation by firm size
Size
JV
M&A
Lic
Contract
Informal
Market
Market
only
Large (>1000)
53
18
21
21
29
47
24
Medium (1001000)
Small (<100)
66
13
10
17
27
36
19
60
8
13
23
41
36
14
All
61
10
13
21
37
37
16
 Small firms favor
informal channels.
 Big firms favor
licensing and M&A
 Medium firms favor
cooperative channels
 Customer-inventions favor
informal and cooperative
channels
 “Specialists” favor market
channels
 Market channels usually
involve JV/ Cooperative R&D
as well
18
Value of inventions by source
 Are high-value “lead user” inventions typical?
 Recall: Customers more frequent
 But incidence ≠value.
 Are innovations from customers more
incremental? or Cheaper to acquire?
 Are lead users typical of customer-sourced
innovation?
 Important details
 Universities, independent inventors and R&D service
providers/consultants are “technology specialists”
 Reference category = pure internal innovations
19
 Higher cost source 
lower share, higher
average sales
 Customer – high
Frequency
Distribution of value of
innovation conditional on the
invention being acquired
incidence, low average
value
 Specialists – lower
incidence, mean high value
 Tests
1.
Value from
customers
2.
Value from
specialists
Cost of acquisition
ccust
tail distribution
unaffected by truncation
Formally correct for
“selection” using MNL
framework
cspe
Value = Vc
20
Frequency distribution of customer
and specialist sourced innovation by
% sales from the focal innovation
160
149
Frequency (weighted)
 Innovations with < 10% of
Specialist
120
revenue: Customers 1.9X
more frequent than
specialist
 Customer innovations are
probably less valuable
Customer
100
80
71
60
36
40
43
42
20
0
Innovations with > 50% of
revenue: Specialist are 2.5X
more frequent than customers
140
80
 Tail Distribution Test:
30
12
<10%
11-25%
26-50%
% sales in 2009 from focal innovation
>50%
Comparing the tails of distribution of observed %
sales from focal innovation
21
Value of inventions by source: MNL approach
with sample selection correction
% firm sales
from focal
innovation
% firm sales from
% firm sales from
focal innovation
focal innovation
(Selection
greater than 50% = 1
correction)
(% firm sales from
focal innovation
greater than 50%) = 1
(Selection correction)
Customer
-3.29** (1.40) -0.07** (0.02)
-3.21** (1.48)
-0.07** (0.02)
Supplier
0.82 (1.79)
0.07** (0.03)
0.97 (1.99)
0.06** (0.03)
Other Firm
3.01 (2.19)
-0.02
3.20 (2.48)
-0.04
Specialists
6.22** (1.64)
0.08** (0.03)
6.30** (1.71)
0.07** (0.03)
Ln (Empl)
-4.10** (1.07) -0.06** (0.02)
-4.10** (1.07)
-0.06** (0.02)
Ind. FE’s (45)
Yes
Yes
Yes
Yes
Controls
Parent size,
Age
Parent size,
Age
Parent size,
Age
Parent size,
Age
-0.21 (1.21)
-0.01 (0.02)
Ln (share of
source in ind)
(0.03)
N
930
930
930
930
R2
0.21
0.15
0.21
0.15
(0.04)
22
Other indicator of innovation value, by source:
Selection corrected estimates
Innovator
Innovator
invests equip or invests in sales
skills
channel
Firm
increased
market share
=1
Firm has
patented
innovation
Customer
-0.03 (0.04)
-0.008 (0.04)
-0.06* (0.04)
-0.10** (0.03)
Supplier
-0.19** (0.05)
0.01 (0.04)
0.07 (0.05)
-0.10** (0.05)
Other Firm
0.01 (0.06)
0.04 (0.06)
-0.05 (0.06)
Specialists
0.14** (0.05)
0.10** (0.04)
0.14** (0.05)
Ln (Empl)
0.06** (0.03)
0.05* (0.03)
0.04 (0.03)
Ind. FE’s
45
45
45
YES
Controls
Ln (share of
source)
N
R2
Parent size, Age
Parent size, Age
Parent size, Age
Parent firm size, Age
-0.08** (0.03)
-0.02 (0.03)
-0.03 (0.04)
1015
0.14
1020
0.16
919
0.13
-0.08 (0.06)
0.29** (0.04)
0.11** (0.03)
0.03 (0.03)
1022
0.25
23
Interpretation
 The value of customer-sourced inventions is lower
than specialist-sourced inventions, but costs of
commercialization and acquisition are also lower.
 Costs: economic proximity reduces search and
contracting costs; lower sales and marketing costs
 Value: inventions incremental,
 Industrial customers disinclined to the changes in existing
equipment, personnel, or even organizations required by more
significant invention by the focal firm (i.e., supplier).
 Specialists—Higher value but higher cost
 Search and contracting costs  higher cost
 not tied to existing products  higher value
24
Conclusions
 Reliance on external sources is high
 Sources of external invention from within the industrial




chain (suppliers, customers) are important in all industries;
“Specialists” disproportionately more in hi-tech industries
Collaboration is major channel for acquiring invention;
Market-based channels (e.g., licensing, M&A) are more
relevant to high-tech industries.
External sources: Customer inventions are low value relative
to specialists, but even lower cost.
High reported patent propensity by innovators
Patents also used by sources, especially specialists and used
in market channels.
25
Thank you
26